Mobile Advertisers, Don’t Be Fooled By Claims About ‘Precise’ Data

Location advertising has captured the imaginations of media creatives and buyers alike, with the promise of precisely targeting the right consumers at the right time and tracking whether marketing drives visitation behavior. Unfortunately, many location data vendors have woven a potentially misleading narrative about the precision of their data.

Data providers often use the words ‘precise’ and ‘accurate’ interchangeably, often boasting that their data has six decimal places as proof of the specificity of their lat/long location fixes. The reality, however, is that these types of claims misrepresent the accuracy of the data. For example, six decimal places in lat/long coordinates actually represents a range of about four inches. The truth is, mobile technology can’t determine a person’s location within four inches, and that measure of precision may not even be relevant when it comes to data accuracy.

Whether it’s intentional or not, the conversation about decimal places and precision has become a distraction from the more important measure of data quality: accuracy.

‘Precision’ and ‘accuracy’ mean two very different things. Lat/long coordinates can be very precise but wildly inaccurate. Precision has to do with specificity, while accuracy is about correctness. Let’s say your friend texts you that she’s standing beside the third pillar on the left, in front of St. Peter’s Basilica in the Vatican. That’s a precise location, but it’s inaccurate if your friend is actually texting you from her desk in San Francisco. If she say she’s in the United States, that is 100% accurate, but it’s also very imprecise. One has nothing to do with the other. When it comes to mobile data, the distinction between accuracy and precision could mean the difference between campaign success and failure.

To be clear, specificity, or precision, is critical. Some of the most valuable use cases for location data require precision. If you want to know that a given device visited a given store, for the purpose of segmenting that user or attributing a visit to marketing investments, you need data that is precise.

Still, precise data isn’t necessarily accurate. Even at its most precise, mobile location data is intermittent, and it’s certainly not perfectly reliable, as some vendors would have you believe. GPS data is the most precise location data that’s widely available for advertising, but a really good way to understand the relation between precision and accuracy is to compare GPS data to another dataset that is more reliable in terms of location accuracy.

When it comes to accuracy, cell tower data is the purity standard. Cell towers can have ranges of multiple miles in areas where cell tower density is low (rural, for instance), so precision isn’t necessarily always high. However, cell tower data is accurate. In fact, cell tower data is on average ten times more persistent and complete than GPS data. It has to be: when someone calls you on your cell phone, your phone rings within a second. That’s only possible because your mobile network provider is constantly interacting with your phone to make sure it knows where to reach you when you get a call. Have you ever noticed how quickly your phone battery drains when you’re on an airplane (and you forget to put your device in “airplane” mode)? That happens because your phone is continually looking for cell towers, trying to maintain constant location communication. There is no “spoofing” of cell tower data. It is the most persistent, complete, and accurate mobile data collected. Wireless carriers would never rely on a spotty “last known location” fix from GPS data to deliver a phone call to your device.

If you were to compare GPS data with cell tower data, you’d find that 30-60% of GPS data points are inaccurate because they don’t match to the possible range of the cell tower(s) to which the device was connected at the time. You’d discover that the GPS location data is often very precise. While it often conveys location coordinates with a very high degree of specificity, it’s frequently incorrect—that’s true of even the most well-known SDKs.

Here’s the critical takeaway: when selecting data for mobile campaigns, don’t base your decision solely on claims about precision or how many decimal places appear in the coordinates, and definitely don’t mistake precision for accuracy. Precision is important, but the value of precision hinges on data accuracy. Don’t take GPS data at face value. It may look really precise, but it can literally be off by miles or far more.

Whenever you’re using location data, make sure to confirm both the precision and the accuracy of the data. Accuracy and precision are not the same thing.

Jake Moskowitz is Head of the Emodo Institute, a dedicated organization within Ericsson Emodo wholly focused on the research, education, and resolution of data concerns that mobile advertisers face.